At 06:32 PM 11/20/00 -0800, Mark Nottingham wrote:
>* Motivation - The draft motivates itself as solving problems in dynamic
> content caching. While I don't want to start nit-picking at this stage,
> IMHO it's important to properly motivate the work, and an invalidation
> protocol doesn't address this issue; it addresses cache coherence where
> resource freshness isn't predictable.
I'd say the motivation for WCIP is two fold, for dynamic content and for
vertical peering. Dynamic content refers to a list of things (frequently
changing, personalized, and dynamically generated from databases). The
current "ObjectList" DTD covers only case 1 (frequently changing ones), but
the trajectory is to carry more complex event notifications for
personalization and dynamically generated stuff. By vertical peering, I
refer to things like preloading, delta update, and log reporting, etc. This
is mainly leveraging the notification transport (the channel and the relay).
>
>* Variants - We'll need to very carefully specify how object variants should
> be handled; in some cases, it will be desireable to invalidate a limited
> selection of variants (e.g., english and german but not japanese), while in
> others it should be possible to invalidate all variants of a particular
> resource.
>
> This should be probably addressed in the channel description format and/or
> the invalidation message format.
Agreed. At first glance (read:don't take my word seriously), the origin
server and/or the invalidation server can set different ObjectNames to
denote different variants (note that invalidations are done based on
ObjectNames), and then we may use regular expression matching for
invalidating a bunch of them. Another approach would be the cache just has
to be aware of all the HTTP variant fields and the invalidation DTD has to
provide fields for those variants. E.g., do a match on all HTTP entity
headers, ... I hope we don't complicate things too much.
>* Transport - WCIP seems to use a hodge-podge of transports, including HTTP,
> a yet-to-be-described TCP protocol, and multicast. I'd argue that there
> should be one transport, or at most two, if multicast is really necessary.
I think "transport" deserves more discussion. There are really two
components in WCIP, 1. a notification transport (including the concept of
channel, relay point, and appl-layer multicast), and 2. the
semantics/messages carried by this transport (right now is ObjectList, but
will expand). Along this line, there's plenty of room to clearly separate
this two and allow growth (separately) in both of them. How Beep (formerly
BXXP) fits in here? a foundation for building WCIP transport?
regards,
Dan
> I'm personally doubtful about using multicast, both because of its poor
> deployment and the fact that it requires specifying two separate models
> for using WCIP. It should be noted that I'm not exactly familiar with the
> issues in multicast, but then again IETF is full of people who are, so
> hopefully this issue can be sorted soon.
>
> I'd very much like BXXP to be addedd to the list of possible components to
> reuse (in design issues). It's a much more functional, clean and
> extensible framework than HTTP, from what little I know.
>
>Cheers,
>
>
>
>
>
>On Wed, Nov 15, 2000 at 12:57:41PM -0800, Dan Li wrote:
> > Hi, please take a look of the WCIP draft before the IETF meeting, as we'd
> > spend time on issue discussions and won't cover the draft in detail. This
> > draft is a strawman, at least the following things need more work:
> >
> > - clearly define the relationship of wcip with normal cache-control
> directives.
> >
> > - specify in more detail the role of relay point, especially for channel
> > aggregation. Can we or can we not aggregate heartbeats, possibly but
> > cutting freshness guarantee.
> >
> > - pick a security model. Use SSL, or PGP signature, or both, and also the
> > security of the channel info.
> >
> > - do we need a default port number? pros and cons wrt. to port 80.
> >
> > - are people happy with persistent TCP connection? are people happy with
> > "delta consistency" (vs. eventual consistency, strong consistency, or
> > anything else).
> >
> > - in the absence of a off-the-shelf realtime reliable multicast protocol,
> > can we still run wcip on multicast UDP for scalability. See section 3.2.3.
> >
> > - preloading, log reporting, advanced event notification, programability,
> > ... future stuff on vertical peering and dynamic content
> >
> > What do people think?
> >
> > Dan
> >
> >
> >
> >
> >
> > At 08:12 AM 11/15/00 -0800, Ian Cooper wrote:
> >
> > >>To: IETF-Announce: ;
> > >>From: Internet-Drafts@ietf.org
> > >>Subject: I-D ACTION:draft-danli-wrec-wcip-00.txt
> > >>Date: Wed, 15 Nov 2000 06:41:12 -0500
> > >>
> > >>A New Internet-Draft is available from the on-line Internet-Drafts
> > >>directories.
> > >>
> > >>
> > >> Title : WCIP: Web Cache Invalidation Protocol
> > >> Author(s) : D. Li, P. Cao, M. Dahlin
> > >> Filename : draft-danli-wrec-wcip-00.txt
> > >> Pages : 26
> > >> Date : 14-Nov-00
> > >>
> > >>Cache consistency is a major impediment to scalable content
> > >>delivery. This document describes the Web Cache Invalidation
> > >>Protocol (WCIP) which uses invalidations and updates to keep
> > >>changing objects up to date in web caches. Moreover, it allows
> > >>automatic one-to-may relay and many-to-one aggregation in a CDN
> > >>(content delivery network) environment.
> > >>
> > >>A URL for this Internet-Draft is:
> > >>http://www.ietf.org/internet-drafts/draft-danli-wrec-wcip-00.txt
> > >>
> > >>Internet-Drafts are also available by anonymous FTP. Login with the
> username
> > >>"anonymous" and a password of your e-mail address. After logging in,
> > >>type "cd internet-drafts" and then
> > >> "get draft-danli-wrec-wcip-00.txt".
> > >>
> > >>A list of Internet-Drafts directories can be found in
> > >>http://www.ietf.org/shadow.html
> > >>or ftp://ftp.ietf.org/ietf/1shadow-sites.txt
> > >>
> > >>
> > >>Internet-Drafts can also be obtained by e-mail.
> > >>
> > >>Send a message to:
> > >> mailserv@ietf.org.
> > >>In the body type:
> > >> "FILE /internet-drafts/draft-danli-wrec-wcip-00.txt".
> > >>
> > >>NOTE: The mail server at ietf.org can return the document in
> > >> MIME-encoded form by using the "mpack" utility. To use this
> > >> feature, insert the command "ENCODING mime" before the "FILE"
> > >> command. To decode the response(s), you will need "munpack" or
> > >> a MIME-compliant mail reader. Different MIME-compliant mail
> readers
> > >> exhibit different behavior, especially when dealing with
> > >> "multipart" MIME messages (i.e. documents which have been split
> > >> up into multiple messages), so check your local documentation on
> > >> how to manipulate these messages.
> > >>
> > >>
> > >>Below is the data which will enable a MIME compliant mail reader
> > >>implementation to automatically retrieve the ASCII version of the
> > >>Internet-Draft.
> > >>
> > >>Content-Type: text/plain
> > >>Content-ID: <20001114131241.I-D@ietf.org>
> > >>
> > >>ENCODING mime
> > >>FILE /internet-drafts/draft-danli-wrec-wcip-00.txt
> > >>
> > >><ftp://ftp.ietf.org/internet-drafts/draft-danli-wrec-wcip-00.txt>
> > >
>
>--
>Mark Nottingham, Research Scientist
>Akamai Technologies (San Mateo, CA)
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:29 MST